May 4th, 2005

Computer History Stuff...

I've let my interest in computer history stagnate for a while, and should probably do something with it. There are so many things people just don't realize about this industry, things that contradict their basic assumptions.

Take Windows. The success of Windows 3.1 wasn't the result of Microsoft "plugging away at it until they got it right", it was the result of one guy working on his own and inventing "Thunking", which suddenly made an obsolete, repeatedly failed back-burner project take off, and resulted in MIcrosoft doing sudden 180 degree reversal in their strategy and breaking up their alliance with IBM.

Or take the Free Software Foundation. People seem to think the FSF invented the idea of free software, but in reality the meanace of proprietary software that it fought against didn't exist until about the time the FSF was formed. Binary code wasn't confirmed to be covered by copyright law until the Apple vs Franklin decision in 1983. That's when the proprietary software industry took off, and only after that did AT&T allow itself to be broken up and really start making Unix proprietary. (Gates really didn't have any legal recourse over the copied paper tapes of Altair basic. He'd have had a hard time suing people even if he'd found out who did what.)

Also, the unit volumes on old computers didn't simply didn't support a retail software industry. Back in the mainframe and minicomputer days, DEC was the leading minicomputer company (and the inventor of the term "minicomputer"). Since minicomputers had higher unit volume than mainframes, it's likely DEC sold more computers than anybody else. (Although since each mainframe cost significantly more than a minicomputer did, IBM made considerably more money than DEC did.)

The DEC PDP8 was introduced in 1965, and was the best selling computer in the world by 1973. As far as anyone can tell it remained the best selling computer until it was displaced by the Apple II (about 5 years later). In its entire history, the PDP8 sold a cumulative total of about 50,000 units. According to The Innovator's Dilemma, in its first two years the Apple II sold 43,000 units, and that was just the start...

When your entire software market is only a few hundred customers, most software is a custom job anyway. The only software out there was either shipped with the hardware, written by the users, or a bespoke job contracted by a user who didn't want to write it themselves. Piracy isn't that big of a problem with bespoke works that need to be installed by professionals over a two-week period. But when you suddenly have millions of potential customers (and microcomputers had unit volume: Sears followed up its sales of the Atari 2600 by selling the Commodore 64), then shrink-wrap binary software becomes commercially viable. The Apple II shipped in 1977, and 1979 saw both the founding of Infocom (producers of the Zork series) and the release of Alkabeth (the first game from Richard Garriott, who went on to found Origin Systems and publish the Ultima series). Wordperfect corporation was also founded in 1979 (under the name SSI). Infocom, Origin, and WordPerfect were some of the first movers in the new retail software industry. It simply didn't EXIST before then. (Before that there was system software bundled with the hardware, which is how Micro-soft started with BASIC and Gary Kildall with CP/M. But the mainframe world had that too, and it wasn't a retail business but an OEM component: as an independent vendor your only customer was the hardware manufacturer.)

So the point is, 1983 is when proprietary software emerged as a driving force, spurred on by the microcomputer industry. (Which became the PC industry after Compaq successfully cloned the IBM PC and _survived_ the legal challenge from IBM becaue they'd been smart enough to use clean-room techniques. The Apple II could easily have been the basis for the PC industry, it had expansion slots just like a PC and could easily have been upgraded to more bits just like the PC was (8086->80386, anyone?). But Apple won the lawsuit against franklin, and thus doomed Apple II clones. IBM lost, and thus the PC industry flourished...

Of course, there's more to it than that. The S/100 systems IBM was cloning were an open standard fromthe start because MITS never had the resources to fight off cloners like Imsai, and went out of business after only a couple years. By making a 16 bit upgrade of an open S-100 system out of commodity components (up to and including the CP/M-86 operating system, whih DOS 1.0 was a bug-for-bug compatible clone of), IBM was positioning the PC to be very very easy to clone.

(And no, Gates didn't show grate foresight for keeping the rights to resell DOS. Those were the same terms Gary Kildall was offering CP/M under. CP/M was a standard part of commodity S-100 systems which had multiple flourishing whitebox OEMs by about 1977. The terms Gates and IBM agreed on were an established standard in that industry niche. Gates sold a similar product, BASIC, which was licensed to dozens of OEMs as well. When Microsoft's first sugar daddy (Mits) went under, the MS remora attached itself to Tandy, and from there to IBM. Microsoft's basic is in the ROM of the commodore 64. Wikipedia has a list.)

Ahem. 1983. The FSF was founded as a conservative "hold the line" entity, fighting against changes in the industry brought on by money and popularity. The goal of the FSF was to preserve the past, the glory days of the 1970's. People forget that. They also forget that the proprietary software industry only established itself 22 years ago, and the first hiccup of it (the founding of Microsoft, the first software company motivated by money rather than technology) was 30 years ago. Yet the first prototype computer system (the Harvard Mark I) was 65 years ago, and the first computer manufactured for sale (the Univac) was 55 years ago. Computers are older than most of the people working with them today, but proprietary software _isn't_.

Open Source is not a new development, proprietary software is. Once you know the history, you can see the real possibility that proprietary software will be looked back upon as an experiment that didn't work.

Too out-there for you? Look at it this way. Yes, enormous fortunes were made. The same is true of the railroad robber barrons over a century ago. There aren't a whole lot of new railroads being built today, instead people drive cars and trucks (steering themselves individually), and fly in airplanes (which can go from any airport to any airport without needing infrastructure built in a fixed route along each possible path). The railroads that are left survive by being a cheap way to haul bulk freight along established high-traffic routes. They wouldn't be cheap if they had to lay new track; they're milking existing infrastructure paid for decades ago simply to survive.

Ten years ago people were pointing out that Microsoft's monopoly was starving investment in new start-up proprietary software firms, and the software industry was stagnating. Open source hasn't made things easier for new start-up proprietary software firms, any more than the automobile made things easier for new start-up railroad companies. With the possible exception of the gaming industry, it's quite possible that the current big players in proprietary software are all we're going to get.

And that's why fifty years from now, proprietary software may be seen as an experiment that didn't work.

Rob
  • Current Mood
    geeky geeky